JP2015049542A - Distributed database system and selection unit - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable a client to access data without being conscious of a storage device that stores the data.SOLUTION: A distributed database system includes a plurality of storage devices 10 and a selection device 20. The storage device 10 has a function of communicating via a telecommunication line 30, and forms a group for each type of stored data. The selection device 20 accesses required data which is specified by identification information notified via the telecommunication line 30. The identification information includes first identification information for identifying the group and second identification information for identifying a storage location of data in the group. The selection device 20 includes: a first extraction unit 21 that accesses the group where the required data is stored, by using the first identification information; and a second extraction unit 22 that accesses the required data by using the second identification information in the group which is accessed by the first extraction unit 21.

Description

  The present invention relates to a distributed database system including a plurality of storage devices and a selection device used therefor.

  In recent years, a distributed database that can be handled by a user as if it were a database using one storage device by integrating a plurality of storage devices communicating via a telecommunication line with a database management system. Technology is adopted.

  The technology described in Patent Document 1 is a management table in which, when data managed by a plurality of resources in a plurality of servers is requested, the requested data is classified in the order of servers, resources, and items by data access processing. Have created. Further, the data access process collects data based on the management table and arranges the data in the order of request so that the data requested by the client can be collectively displayed.

JP 2000-57074 A

  The technique described in Patent Document 1 solves the problem of enabling the client to acquire data in a batch without being aware of the data arrangement or the like. That is, the data access process performs a process of collecting data managed by being distributed to a plurality of servers and a process of arranging the data in order.

  However, in this configuration, the client must specify the server that manages the data in order to instruct the requested data to the data access process, and if the client does not know which data exists on which server There is a problem that data cannot be accessed.

  An object of the present invention is to provide a distributed database system that enables a client to access data without being aware of a storage device storing the data, and further provides a selection device used in the distributed database system. For the purpose.

  The distributed database system according to the present invention has a function of communicating through a telecommunication line, a plurality of storage devices forming a group for each type of data to be stored, and identification information notified through the telecommunication line And a selection device that accesses the required data specified in (1), wherein the identification information includes first identification information that identifies the group, and second identification that identifies a storage location of the data in the group And the selection device uses the first identification information to access the group storing the required data and the group accessed by the first extraction unit. And a second extraction unit that accesses the required data using the second identification information.

  In this distributed database system, it is preferable that the selection device is a server that communicates with the storage device through the electric communication line.

  The selection device according to the present invention is the server used in the distributed database system described above.

  According to the configuration of the present invention, a group is formed for each type of data stored in a plurality of storage devices, and the storage devices storing the data are narrowed down using the first identification information for identifying the group. Yes. Furthermore, using the second identification information for identifying the data storage location in the group, the required data is accessed in the group. Therefore, the client can access the data without being aware of the storage device storing the data.

It is a block diagram which shows embodiment. It is a schematic block diagram which shows one example of use same as the above. It is a schematic block diagram which shows the other usage example same as the above. It is the schematic which shows the structural example of the memory | storage device same as the above.

  A distributed database system (hereinafter abbreviated as “distributed database”) described below includes a plurality of storage devices 10 and a selection device 20, as shown in FIG. The storage device 10 has a function of communicating through the telecommunication line 30 and forms groups G1 and G2 for each type of data to be stored. The selection device 20 accesses necessary data specified by the identification information notified through the telecommunication line 30. The identification information includes first identification information for identifying a group, and second identification information for identifying a data storage location in the groups G1 and G2. The selection device 20 uses the first identification information to access the first extraction unit 21 that accesses the groups G1 and G2 in which necessary data is stored, and the first extraction unit 21 accesses the groups G1 and G2. And a second extraction unit 22 that accesses the required data using the two identification information.

  The configuration for realizing the storage device 10 may be any configuration such as a dedicated storage such as a hard disk device, a storage handled by the database management system, and a data table as long as data can be stored. However, the storage device 10 must be able to communicate through the telecommunication line 30. Therefore, the storage device 10 is realized as a server composed of a computer system.

  In the groups G1 and G2 formed by the storage device 10, a database server 41 (that is, a server equipped with a database management system) is provided for each of the groups G1 and G2. A plurality of database servers 41 may be provided in the groups G1 and G2. The selection device 20 is realized by a dedicated server provided with an application program interface (Application Programming Interface: API). That is, the selection device 20 is preferably a server that communicates with the storage device 10 through the telecommunication line 30. When the telecommunication line 20 is the Internet, the selection device 20 is configured by a web server. In the illustrated example, only one selection device 20 is shown, but it goes without saying that a plurality of selection devices 20 may be provided as necessary. The API may be provided in the database server 41. In this case, the selection device 20 is provided in the database server 41. Alternatively, an API may be provided in any storage device 10 and used as the selection device 20.

  In the illustrated example, the selection device 20 communicates through the server 42 with the terminal device 43 as a client and the equipment 44 as a management target through the telecommunication line 30. The selection device 20 communicates with the database server 41 through the telecommunication line 30. The server 42 is provided for monitoring the equipment 44 or controlling the equipment 44 through the telecommunication line 30, and data acquired from the equipment 44 by the server 42 is stored in the storage device 10 through the selection device 20. Further, when the terminal device 43 requests data stored in the storage device 10, the server 42 reads out necessary data from the storage device 10 through the selection device 20. That is, the selection device 20 has a function of accessing data in the storage device 10 in response to a request from the server 42. Here, accessing data means both retrieving data stored in the storage device 10 and storing data in a required location of the storage device 10.

  In FIG. 1, the telecommunications line 30 to which the server 42 is connected and the telecommunications line 30 to which the selection device 20 is connected are described separately because they have different data contents. The line 30 may be physically the same. Further, when the server 42 mainly collects data, it is desirable that the server 42 and the equipment 44 use M2M (Machine-to-Machine) technology.

  In the configuration example shown in FIG. 1, the distributed database can be shared with the server 42 that provides various services by separating the server 42 and the selection device 20, but the selection device 20 is shared by the server 42. It is also possible to adopt a configuration that does this. The function of the server 42 will be described later. The telecommunication line 30 is assumed to be a VPN (Virtual Private Network) using the Internet, but may be the telecommunication line 30 using a dedicated communication path.

  Assume that a distributed database is used to handle more than 4 million data on electricity, gas, water, etc. of 150,000 facilities. When automatically collecting data of this scale, for example, it is necessary to secure the performance of processing data with a total data capacity of 20 terabytes or more for 3 million points or more. Examples of facilities include business buildings (such as office buildings and commercial facilities), residential buildings (such as apartment buildings), parks, and exercise facilities. The data preferably includes data relating to the usage of electricity, gas, water, etc. (hereinafter referred to as “energy data”), data relating to the configuration of equipment provided in the facility, data relating to the operating state of the equipment, and the like. The period for automatically collecting data is required to be shorter and is currently about 10 minutes, but it is also required to be shortened to 1 minute or less.

  As mentioned above, in addition to energy data, if various data such as the configuration of equipment and the operating state of equipment can be managed in an integrated manner, by combining these data in various ways and processing the data, It becomes possible to provide various services. For example, various services such as a service that collects energy data, a service that visualizes the transition of energy data, a service that manages the total amount of power, gas, and water used in the facility, and a service that monitors the operating status of equipment provided in the facility Service becomes available.

  The type of service provided using the distributed database is determined by a contract with the user who receives the service, and the server 42 provides the service determined by the contract with the user. Since the service is provided through the telecommunications line 30, the area where the service can be provided is wide, and the service can be provided not only in Japan but also overseas. Therefore, it is desirable that the distributed database has a cloud.

  In addition, since the type of data to be used varies depending on the type of service provided to the user, it is desirable to divide the storage device 10 that stores data according to the type of data. In other words, the amount of information varies depending on the type of data, so when multiple types of data are mixed, handling the data in a batch increases the amount of information depending on the type of data contained in the batched data. Variation occurs. Therefore, if this type of data is stored in one storage device 10 in a state where a plurality of types of data are mixed, the access time of the storage device 10 varies greatly.

  In the present embodiment, the storage device 10 forms groups G1 and G2 for each type of data to be stored. As described above, when there are three types of data, energy data, equipment configuration data, and equipment operating state data, three groups of storage devices 10 are provided, and each type of data is stored in different groups. Is done. By forming such groups G1 and G2, variation in access time of the storage device 10 is suppressed in one group G1 and G2.

  In the configuration example shown in FIG. 2, the storage device 10 constitutes three groups: a group G1 that stores master management data, a group G2 that stores energy data, and a group G3 that stores facility operation data. The master management data is data relating to the configuration of equipment, and stores, for example, the hierarchical structure of equipment such as a user (business) -business establishment-building-floor-equipment, equipment specifications, and the like.

  The configuration example illustrated in FIG. 2 illustrates a case where various servers 42 perform various services by using the above-described distributed database system (cloud) as a common platform. The upper server 42 mainly provides a service for providing data (information) to the terminal device 43, and the lower server 42 mainly performs a service for collecting data from the equipment 44.

  In the illustrated example, the lower server 42 provides energy collection, BEMS (Building and Energy Management System) monitoring, cooling / air conditioning, and remote monitoring services, and the upper server 42 provides information transmission and total amount management, respectively. It shows the case of providing services for analysis, analysis and diagnosis. BEMS monitoring and remote monitoring are services for remotely monitoring or controlling the equipment 44 using data collected from the equipment 44. BEMS monitoring is a service that supports control for suppressing power consumption of lighting equipment and air conditioning equipment in office buildings and tenant buildings, and remote monitoring is a service that performs remote control of power storage equipment and the like. Refrigeration / air conditioning is a service that performs remote monitoring, etc. for refrigeration / refrigeration showcases and air conditioning equipment in stores. The information transmission is a service for transmitting information such as product information and energy saving cases, and the total amount management is a service for indicating the energy usage amount in a graph or the like. Visualization is a service that shows the operation status of solar power generation equipment, power storage equipment, etc. in a graph or the like. The analysis / diagnosis is a service for performing analysis and diagnosis so that the equipment 44 can be efficiently operated without waste.

  However, the configuration example shown in FIG. 2 is an example, and the type and contents of the service are appropriately selected. In addition, as shown in FIG. 3, in a configuration including a plurality of storage devices 10 that are geographically distributed, one or more types of storage devices 10 may be arranged for each region. . In the illustrated example, three regions D1, D2, and D3 surrounded by an alternate long and short dash line are shown. In the two regions D1 and D3, the group G2 and the group G3 are used in combination, and in the remaining one region D2, only the group G2 is used. ing. The storage device 10 of the group G3 in the region D1 at the left end of the figure stores, for example, facility operation data relating to the operation state of the photovoltaic power generation facility and the power storage facility. In addition, the storage device 10 of the group G3 in the region D3 on the right end of the figure stores, for example, facility operation data relating to the operating states of the store freezer and the refrigerator.

  As in the example illustrated in FIG. 3, the storage devices 10 of the same group G1, G2, G3 may be provided in different areas. For example, the storage devices 10 belonging to the group G2 that stores energy data may be provided in Tokyo and Osaka, respectively. A plurality of storage devices 10 belonging to the same group G1, G2, G3 may be arranged in the same area.

  On the other hand, when data necessary for providing a service is distributed and stored in a plurality of storage devices 10, the server 42 must access data in the plurality of storage devices 10. Even in this case, the user can access the required data from the terminal device 43 or the facility 44 without being aware of the configuration and location of the storage device 10 in which the data is stored. That is, even when accessing data stored in a plurality of storage devices 10, access to required data from the terminal device 43 or the facility 44 is as if accessing one database server. To be done seamlessly.

  The selection device 20 performs access to data necessary for the user to receive the service. That is, when receiving a request from the server 42, the selection device 20 accesses the storage device 10 according to the type of service. That is, a user receives a service through the server 42, and the server 42 accesses data according to the service provided by the server 42 by using a distributed database constructed as a cloud. Since the selection device 20 is notified of identification information corresponding to the type of service from the server 42, the selection device 20 uses the first identification information and the second identification information included in the identification information as described above. To access the required data.

  2 and 3, the selection device 20 is provided across a plurality of groups G1, G2, G3 or a plurality of regions D1, D2, D3, but the selection device 20 is a distributed database as a cloud. Therefore, it may be appropriately divided. That is, as described above, a plurality of selection devices 20 may be included in the distributed database.

  Hereinafter, the identification information is referred to as “global ID”. The global ID further includes “issuer ID” that is first identification information and “local ID” that is second identification information. The issuer ID is determined to be unique within the range of the distributed database, and the local ID is determined to be unique within the range of the issuer ID.

  The issuer ID is information for identifying a group in the storage device 10. In this embodiment, the issuer ID is information for identifying a group G1, G2, G3. Information for identifying the location area and information for identifying the user are combined. Each location and user is given a unique number that serves as identifying information. Similarly, the information for identifying the group is given a unique number for each group. The location is set in units selected from the country, prefecture (region), municipality, and the like. Accordingly, the issuer ID is represented by a set of (country code, user number, group number), for example. Of course, the information for identifying the location, the user, and the group may not be a number.

  On the other hand, since the local ID only needs to be set so as to be unique among the issuer IDs, information indicating the order within the range of the local ID is usually given. That is, as the local ID, a number indicating the order in which data is stored in the storage device 10 is used. When the local ID is used, the location where the data is stored in the group is specified. Therefore, the data can be stored in the location and the data stored in the location can be accessed. In other words, it can be said that the local ID identifies data in the group. The reason for using a number representing the order for the local ID will be described later.

  When the selection device 20 receives the global ID, the first extraction unit 21 specifies the group of the storage device 10 using the issuer ID. That is, when the selection device 20 is requested to access the data, the first extraction unit 21 identifies the group of the storage device 10 using the country code, the user number, and the group number. Narrow down the storage device 10 as a target. Thereafter, the second extraction unit 22 uses the local ID to identify a location where the required data is stored in the group and accesses the required data.

  As described above, the selection device 20 specifies a location where necessary data is stored by two-step processing using the first extraction unit 21 and the second extraction unit 22, and accesses target data. Will do.

  By the way, if an attempt is made to provide such a service, a large amount of data is generated at a relatively short time interval. Therefore, input / output for data collection and provision becomes a bottleneck, and exclusive control is performed. Increases frequency. To deal with this type of problem, it is effective to use a distributed database. In a distributed database, data is distributed and stored in a plurality of storage devices 10, so that it is possible to avoid a bottleneck in a database constructed with one storage device 10.

  However, when there are a plurality of storage devices 10 as physical entities that construct a distributed database, storage devices 10 with different hardware resources may coexist. In the database, among the hardware resources, the difference between the performance of the CPU (Central Processing Unit) and the storage capacity greatly affects the throughput. Therefore, due to variations in hardware resources of the storage devices 10 constituting the distributed database, variations occur in writing time when storing data and response time when providing data.

  Further, when data is stored in a plurality of storage devices 10 constituting a distributed database, access may be concentrated on a single storage device 10, and in this case, the corresponding storage device 10 becomes a bottleneck. Become. That is, the overall throughput of the distributed database is reduced.

  The present embodiment employs the following configuration so that the occurrence of bottlenecks is suppressed even if there are variations in hardware resources among the plurality of storage devices 10 constituting the distributed database. That is, by adopting the following configuration, it is possible to provide a distributed database that is unlikely to cause a reduction in throughput even when storage devices 10 with inferior hardware resources are mixed.

  As described above, the distributed database includes a plurality of storage devices 10, and the distributed database and the terminal device 43 communicate with each other through the electric communication line 30 to construct a client server system using the terminal device 43 as a client. The client server system has a three-layer architecture including a presentation layer, an application layer, and a data layer, and the presentation layer is realized by a terminal device 43.

  The storage device 10 corresponds to the data layer, and the selection device 20, the database server 41, and the server 42 correspond to the application layer. The server 42 has a function of making a request corresponding to the processing requested from the terminal device 43 to the database server 41 and returning a response of the database server 41 to the terminal device 43.

  As shown in FIG. 4, the storage device 10 is divided into a plurality of platform modules (hereinafter referred to as “PF modules”) 11 as virtual servers. The selection device 20 has a function of integrating a plurality of PF modules 11 as well as a function of handling a plurality of storage devices 10 as one database. This function is realized by an API provided in the selection device 20 as in the case of a function that integrates a plurality of storage devices 10 and handles them as one database. That is, the API of the selection device 20 includes a function or an instruction for realizing a function of selecting the PF module 11 corresponding to a request from the terminal device 43.

  When the selection device 20 manages the PF module 11, the PF module 11 functions equivalently to an individual server when viewed from the server 42. That is, each PF module 11 apparently has a function of storing data and a function of a database management system. The individual PF modules 11 are constructed so that the storage capacities are equal. In other words, the PF modules 11 have the same storage capacity regardless of the storage capacity of the actual storage device 10. In other words, it can be said that the distributed database is configured by combining a plurality of PF modules 11 having the same storage capacity.

  Now, in order to suppress the occurrence of a bottleneck, a case is assumed in which a load is distributed using 100 PF modules 11 in a distributed database. When a distributed database is introduced, if only one storage device 10 having a storage capacity of 50 gigabytes is used, 100 PF modules 11 must be provided in one storage device 10, so that one PF The storage capacity of the module 11 is 500 megabytes.

  Here, it is assumed that a 150 gigabyte storage device 10 is newly introduced as the amount of data to be stored increases. The total storage capacity of the two storage devices 10 is 200 gigabytes. Similarly to the case of one storage device 10, when 100 PF modules 11 are configured, the storage capacity of one PF module 11 is Will be 2 gigabytes. Since the storage capacity of the first storage device 10 is 50 gigabytes, 25 PF modules 11 can be configured. Since the storage capacity of the second storage device 10 is 150 gigabytes, 75 PF modules 11 are used. Can be configured.

  Further, it is assumed that the number of storage devices 10 is increased as the amount of data to be stored increases in the same manner as the data to be stored. Here, it is assumed that 100 storage devices 10 having a storage capacity of 50 gigabytes are provided. In this case, since the entire distributed database is 5 terabytes, if 100 PF modules 11 are provided, one PF module 11 has a storage capacity of 50 gigabytes. Further, if 100 PF modules 11 are required, one storage device 10 corresponds to one PF module 11.

  As a general tendency, the storage capacity of the storage device 10 increases as the processing capability increases. Here, since the storage capacities of the PF modules 11 are equal, the throughputs of the PF modules 11 are substantially equal regardless of the processing capability of the storage device 10. Although the storage area of the PF module 11 is further divided, it is not a gist and will not be described here.

  The local ID (second identification information) extracted by the second extraction unit 22 provided in the selection device 20 may be identification information that identifies the PF module 11. In this case, in order to specify the location where the required data is stored, identification information indicating the location where the data is stored in the PF module 11 is required.

  In order to distribute the load so that the data is distributed and stored in a plurality of PF modules 11, for example, a data ID of a natural number including 0 is given to the data, and the data ID is the number of PF modules 11. The remainder when divided by is used as the local ID. That is, if the data ID is X and the number of PF modules 11 is K, the local ID is a remainder obtained by dividing X by K (that is, X mod K). The local ID is a numerical value starting from 0. Therefore, when the remainder becomes 0, the PF module 11 having the local ID 0 is selected.

  In this case, when the data IDs of data stored in order in the storage devices 10 of the same group are different by one, such data is stored in different PF modules 11 in order. Therefore, data is distributed and stored in different PF modules 11 as if striping is performed using a plurality of hard disks. As a result, access is not concentrated on a single PF module 11, and access to data is leveled.

  As described above, the global ID is composed of the issuer ID and the local ID, and the group of the storage device 10 for storing data is obtained using the issuer ID, and then the data using the local ID is obtained. Is required. Therefore, when extracting required data, by specifying the type of data, the group of the storage devices 10 is narrowed down, and the PF module 11 can be specified using the local ID within the group. Eventually, it is only necessary to access data within the range of one PF module 11, so that access to the data is easy.

  In the configuration example described above, in order to determine the PF module 11 for storing data, the selection device 20 performs division of the data ID to obtain a remainder. However, the PF module stores data using another relationship such as a random number. May be dispersed.

  As can be seen from the above-described example, since the PF module 11 is virtually configured using the storage device 10 as an entity, not only the load is leveled, but also the amount of data stored in the storage device 10 is increased. Accordingly, the number of storage devices 10 can be increased in stages. That is, it becomes possible to ensure the scalability of the throughput and storage capacity of the distributed database. The distributed database described in the embodiment can be constructed regardless of whether the equipment to be configured is outsourced to a specialist (housing) or the equipment of the specialist is used (hosting). Therefore, the distributed database can be operated in an environment suitable for the service provided by the server 42.

DESCRIPTION OF SYMBOLS 10 Memory | storage device 20 Selection apparatus 21 1st extraction part 22 2nd extraction part 30 Telecommunication line 41 Database server 42 Server 43 Terminal device 44 Equipment G1, G2 group

Claims (3)

A plurality of storage devices having a function of communicating through a telecommunication line and forming a group for each type of data to be stored;
A selection device for accessing required data specified by identification information notified through the telecommunication line,
The identification information is
First identification information for identifying the group;
Second identification information for identifying the data in the group,
The selection device is:
A first extraction unit that accesses the group in which the required data is stored using the first identification information;
A distributed database system comprising: a second extraction unit that accesses the required data using the second identification information in the group accessed by the first extraction unit. The distributed database system according to claim 1, wherein the selection device is a server that communicates with the storage device through the telecommunication line. It is the said server used for the distributed database system of Claim 2. The selection apparatus characterized by the above-mentioned.

Images